A machine vision cooperative measurement method of multiple field of view (Multi-FOV) was proposed to realize the high-precision automatic measurement of two-dimensional geometric features. The basic principle of the method was introduced and the key technologies to realize the cooperative measurement were studied. Firstly, a measuring space was established, the features to be measured were identified, and an optimized measuring path was planed in a large FOV image. Then, the mapping relation between the image coordinate of the large FOV and the measuring space coordinate was established. Guided by the measuring path in the large FOV image and the control system, the sequential images with small FOVs were collected automatically in the measuring space. Furthermore, the mapping relation between the image coordinates of the large FOV and the small FOVs and that between the adjacent small FOV image coordinates were derived. On the basis of these relationships, exact auxiliary measurement characteristics in the corresponding locations in the small FOV images were researched and constructed. Finally, according to the orientations and positions of sequential small FOV images in the measuring space, the partial parameters were calculated and then further add them to get the whole parameters of the measured features. Experimental results indicate that the absolute value of relative error is less than 0.03% when the method was used to gauge the distance of circular holes with the nominal dimension of 100 mm. The outstanding advantage of the cooperative measurement method of multi-FOV is that the measuring accuracy is not influenced by ambient temperature and the coordinate precision of machine. It is an effective method to automatically measure two-dimensional complex geometric characteristics of mechanical parts in industrial fields.